Professional Documents
Culture Documents
3, June 2019
ABSTRACT
Optimal components assignment problem subject to system reliability, total lead-time, and total cost
constraints is studied in this paper. The problem is formulated as fuzzy linear problem using fuzzy
membership functions. An approach based on genetic algorithm with fuzzy optimization to sole the
presented problem. The optimal solution found by the proposed approach is characterized by maximum
reliability, minimum total cost and minimum total lead-time. The proposed approach is tested on different
examples taken from the literature to illustrate its efficiency in comparison with other previous methods.
KEYWORDS
Components Assignment Problem, Stochastic-Flow Networks, Network Reliability, Fuzzy Multi-Objective
Linear Programming, Genetic Algorithms.
1. INTRODUCTION
Network reliability of stochastic-flow network (SFN) is defined as the probability that a specified
amount of flow can be transmitted successfully from source to destination through SFN [1].
Components assignment problem (CAP) is one important problem in the field of system
reliability analysis, finding an optimal component assignment is significant to maximize the
system reliability and improve the system performance [2]. Many researchers studied CAP for a
SFN to maximize the network reliability under different constraints, [3], proposed an algorithm to
generate all minimal system states fulfilling the demand, time and budget constraints, then the
system reliability is evaluated in terms of such system states. The authors in [4] focused on
finding the optimal carrier selection based on network reliability criterion under a budget
constraint, an optimization algorithm integrating a genetic algorithm, minimal paths and the
recursive sum of disjoint products is proposed to solve such a problem. Multi-state CAP was
discussed in [5] to maximize the network reliability under an assignment budget constraint, in
which each component has an assignment cost, they suggested an optimization method based on
genetic algorithm. In [6] they studied the optimal network line assignment with maximal network
reliability and minimal total coast, they presented an approach based on Non-dominated Sorting
Genetic Algorithm II (NSGA-II) and Technique for Order Preference by Similarity to Ideal
Solution (TOPSIS) to solve multi-objective optimization for stochastic computer networks. In
order to solve multi-objective CAP, [7], proposed two-stage approach to solving the multi-
DOI : 10.5121/ijmnct.2019.9301 1
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
objective CAP subject to reliability and assignment cost for SFN. In [8] he proposed an approach
to get the exact optimal double-resource assignment for the robust design problem in multistate
computer networks, a minimum capacity assignment for each link and node is searched to keep
the network working even both links and nodes are subject to failures.
In the case of CAP for stochastic-flow network under lead-time constraint,[9], discussed this and
he suggested Genetic Algorithm(GA) to search the optimal components for a minimum total lead-
time that maximizes the system reliability, such that the total lead-time cannot exceed a specified
amount. In addition, [10] studied multi-objective CAP subject to lead-time constraint they
proposed GA based on the NSGA-II to search the optimal components that maximize the
reliability. In the case of each component has both an assignment cost and lead -time constraints,
[11], the CAP for SFN was studied and solved by a proposed approach based on a random
weighted GA. The objective of proposed approach was to maximize the network reliability,
minimize total lead–time and minimize cost.
The concept of decision making in the fuzzy environments is presented by [12]. In [13] illustrated
that without increasing the computational effort, Fuzzy Linear Programming( FLP) problems can
be solved. In addition, [14] presented general look at core ideas that make up the burgeoning
body of fuzzy mathematical programming emphasizing the methodological view, and so [15]
aggregated the concept of multi-objective programming application and using a membership
function of the linear expression to represent and integrate each fuzzy objective, he let the
solution is converted to another form of linear programming solution by using the way solve the
application problem of fuzzy theory. Where in[16] they presented an inexact approach and
recommended genetic algorithm to get a family of inexact solutions with acceptable membership
degree to solve objective and resource type of FLP problems. A type of model of fuzzy quadratic
programming problems is proposed in [17], according to different types of fuzzy resource
constraints and fuzzy objective in actual production problems, they described the fuzzy objective
and resource constraints with different type of membership functions. Furthermore, FLP problem
formulations and membership functions were discussed by many researchers, [18 – 31] to apply
FLP to various problems and improve the obtained solutions.
Recently, FLP is used to solve various problems [32-37]. By using a fuzzy multi-objective GA,
[33] succeed in obtaining high quality solutions to solve the multi-objective decision problem.
While in [34] they applied a fuzzy multi-objective linear programming model to combine the
existed components with a new character by using an optimization method of the highest match.
In [35] a new ranking methods of Subinterval average and subinterval addition is presented in
order to solve FLP problem. A fuzzy linear programming model for a problem of food industry is
presented and solved by [36] .The FLP is applied to the tri generation system (power generation,
heat generation, and the generation of cooling effect), [37], to find the optimal design to the
proposed system.
The aim of this paper is to solve the CAP for an SFN under system reliability, total lead time and
total cost constraints. An approach based on fuzzy linear programming is presented to solve the
CAP.
The paper is organized as follows: Section 2 illustrates needed notations. Section 3 presents the
problem formulation. Next, section 4 illustrates the fuzzy linear formulation to the presented
problem. Section 5 explains the proposed multi-objective GA based on fuzzy linear
2
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
programming. To demonstrate the usability of the proposed approach, several examples included
in Section 6. Section 7 presents comparison and discussion, the last section shows conclusion.
2. NOTATIONS
𝑁 No. of nodes.
𝑣 {𝑎𝑒 |1 ≤ 𝑒 ≤ 𝑣} : No. of arcs.
𝑀𝑃𝑠 Minimal paths.
𝑛𝑝 Number of minimal paths.
𝑚𝑝𝑗 Minimal path no. 𝑗; 𝑗 = 1, 2, … , 𝑛𝑝.
𝑣𝑐 The number of available components.
𝑣𝑛𝑘 The components number 𝑘, 𝑘 = 1,2, … , 𝑣𝑐.
𝑙(𝑣𝑛𝑘 ) Lead time of components𝑣𝑛𝑘 .
𝑐(𝑣𝑛𝑘 ) Cost of components𝑣𝑛𝑘 .
𝐿𝑗 The lead time of𝑚𝑝𝑗 .
𝑅𝑑,𝑇 The system reliability to the demand d under time limit 𝑇, for simplicity using 𝑅.
𝓍 Capacity vector defined as 𝒳 = (𝑥1 , 𝑥2 , … . . , 𝑥𝑒 ,).
𝑃 (𝑝1 , 𝑝2 , … , 𝑝𝑣 ) The components assignment in which 𝑣𝑛𝑘 is assigned to the arc 𝑎𝑒 if
𝑝𝑒 = 𝑘.
𝑆𝑙 (𝑃 ) Total lead time.
𝐶 (𝑃 ) Total cost.
𝒮 Population size.
ℊ Maximum number of generations.
𝑔𝑛 Generation number.
𝑔𝑚 GA mutation rate.
𝑔𝑐 GA crossover rate.
𝑆𝑙𝑜𝑏𝑗 Minimum acceptable feasible values of𝑆𝑙 (𝑃 ).
𝑆𝑙0 Maximum acceptable feasible values of𝑆𝑙 (𝑃 ).
𝑅𝑜𝑏𝑗 Maximum acceptable feasible values of 𝑅 .
𝑅0 Minimum acceptable feasible values of 𝑅 .
𝐶𝑜𝑏𝑗 Minimum acceptable feasible values of 𝐶 (𝑃 ).
𝐶0 Maximum acceptable feasible values of 𝐶 (𝑃 ).
𝜇(R) Fuzzy objective membership functions of 𝑅 .
𝜇 (𝑆𝑙 ) Fuzzy objective membership functions of 𝑆𝑙 (𝑃 ).
𝜇 (𝐶 ) Fuzzy objective membership functions of𝐶 (𝑃 ).
α The acceptable membership degree level.
3. PROBLEM FORMULATION
The mathematical programming formulation of the multi-objective optimization problem to
maximize system reliability of a flow network, minimize the total lead-time and cost illustrating
as follow:
3
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
Subject to:
𝑝𝑒 = 𝑘, 𝑘 ∈ {1,2, … , 𝑣𝑐}for e = 1,2, … , v. (4)
𝑝𝑒 ≠ 𝑝ℎ for 𝑒 ≠ ℎ (5)
𝐿𝑗 ≤ 𝑇, j = 1,2, … , np (6)
Where:
𝐿𝑗 = ∑𝑉𝑒=1 𝑙(𝑝𝑒 ) | (7)
𝑝𝑒 ∈ 𝑚𝑝𝑗
𝑉
𝑆𝑙 (𝑃 ) = ∑𝑒=1 𝑙(𝑝𝑒 ) (8)
𝐶 (𝑃 ) = ∑𝑉𝑒=1 𝐶(𝑝𝑒 ) (9)
And, constraints (4) and (5) emphasize that each link should be given one component and that
each component can be assigned to at most one link. All feasible component assignments are
generated using constraints (4) and (5). Constraint (6) assures that the lead-time of the path 𝑀𝑃𝑗
(𝐿𝑗 ) is less than the time limit (𝑇), [9].
𝑜𝑏𝑗
𝑅 ≤ 𝑅𝑜𝑏𝑗 ,𝑆𝑙 (𝑃 ) ≥ 𝑆𝑙 , 𝐶 (𝑃 ) ≥ 𝐶𝑜𝑏𝑗 .
1 if 𝑅 > 𝑅𝑜𝑏𝑗
𝑅𝑜𝑏𝑗 − 𝑅
𝜇(𝑅) = 1− 𝑖𝑓 𝑅𝑜𝑏𝑗 − 𝒫0 ≤ 𝑅 ≤ 𝑅𝑜𝑏𝑗 (10)
𝒫0
{ 0 if 𝑅 < 𝑅0
𝑜𝑏𝑗
1 𝑖𝑓 𝑆𝑙 (𝑃 ) < 𝑆𝑙
𝑜𝑏𝑗
𝑆𝑙 (𝑃 ) − 𝑆𝑙 𝑜𝑏𝑗 𝑜𝑏𝑗
𝜇 (𝑆𝑙 ) = 1 − 𝑖𝑓 𝑆𝑙 ≤ 𝑆𝑙 (𝑃 ) ≤ 𝑆𝑙 + 𝒫1 (11)
𝒫1
{ 0 𝑖𝑓 𝑆𝑙 (𝑃 ) > 𝑆𝑙0
And,
1 𝑖𝑓 𝐶 (𝑃 ) < 𝐶𝑜𝑏𝑗
𝐶 (𝑃 ) − 𝐶𝑜𝑏𝑗
𝜇(𝐶 ) 1− 𝑖𝑓 𝐶𝑜𝑏𝑗 ≤ 𝐶 (𝑃 ) ≤ 𝐶𝑜𝑏𝑗 + 𝒫2 (12)
𝒫2
{ 0 𝑖𝑓 𝐶 (𝑃 ) > 𝐶0
Where:
𝒫0 Tolerance of𝜇(𝑆𝑙 ), 𝒫0 = 𝑅𝑜𝑏𝑗 − 𝑅0 .
𝑜𝑏𝑗
𝒫1 Tolerance of𝜇(𝑅), 𝒫1 = 𝑆𝑙0 − 𝑆𝑙 .
𝒫2 Tolerance of𝜇(𝐶 ), 𝒫2 = 𝐶0 − 𝐶𝑜𝑏𝑗 .
4
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
Hence, the membership function of the decision space 𝑆̅ is 𝜇𝑠̅ (𝑃 )is given by:
Max 𝜇𝑠̅ (𝑃 ) = Max{0, min{𝜇(R), 𝜇 (𝑆𝑙 ), 𝜇(𝐶 )}} (13)
The chromosome 𝑃 contains 𝑣fields, where 𝑣is the number of arcs (components) for the network.
Each field in 𝑐ℎrepresents the components number assigned to an arc.
𝑃 = (𝑝1 , 𝑝2 , … , 𝑝𝑣 )
Where𝑝1 , 𝑝2 𝑎𝑛𝑑 𝑝𝑣 are random component numbers between 1 and𝑣𝑐, this mean that the
component 𝑝1is assigned to arc𝑎1 , the component 𝑝2 is assigned to arc 𝑎2 ,…and the component
𝑝𝑣 is assigned to arc 𝑎𝑣 .
We take the membership function of the fuzzy optimal solution, 𝜇𝑠̅ (𝑃 )as the fitness function 𝐹 of
the genetic algorithm.
We will use the roulette wheel selection method to select the parent population to the next
generation from the current population as follow:
Step 1: calculate a cumulative probability for each chromosome 𝑝𝑟(𝑔𝑛), 𝑔𝑛 = 1,2, … , 𝒮 by:
𝜇 (𝑃)
𝑝𝑟(𝑔𝑛) = 𝒮 𝑠̅
∑𝑔𝑛=1 𝜇𝑠̅ (𝑃)+ 𝜀
(14)
Where𝜀 is small positive integer, it used to guarantee a nonzero denominator.
Step 2: generate random real number 𝑟 in [0, 1].
Step 3: if r ≤ 𝑝𝑟(1) , select the first chromosome, otherwise select the 𝑔𝑛𝑡ℎ chromosome
(2 ≤ 𝑔𝑛 ≤ 𝒮 ) 𝑖𝑓 𝑝𝑟(𝑔𝑛 − 1) < 𝑟 ≤ 𝑝𝑟(𝑔𝑛).
Step 4: Repeat steps 2 and 3, 𝒮 times and obtain 𝒮 chromosomes.
5
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
In the proposed GA, uniform crossover is used to breed a child from two parents by randomly
taking a component from the corresponding component of the child as shown in fig.1.The
crossover operation is performed as follows:
Step 1: select two chromosome according to the selection strategy, section 5.4.
Step 2: randomly take a component from one of the two chromosomes to form a corresponding
components of the child.
Step3: repeat step 2 until the components of the child fill up perfectly.
A child undergoes mutation according to the mutation probability 𝑔𝑚 and the mutation
probability for each component𝑔𝑚 .
Step 1: generate a random number 𝑟1 ∈ [0,1].
Step 2: if𝑟1 < 𝑔𝑚 , the chromosome is chosen to mutate and go to step 3, otherwise skip this
chromosome.
Step 3: for each component of the child do:
Step 3.1: Generate a random number 𝑟2 ∈ [0,1].
Step 3.2: if 𝑟2 < 𝑔𝑚 then mutate this component as follows:
Step 3.2.1: if𝑝𝑗 = 𝑣𝑛𝑘 , then randomly choose one in {1,2, … , 𝑣𝑐} − {𝑣𝑛𝑘 }.
Step 3.2.2:if previous step does not achieve skip this component.
Figure 2 shows an example of performing the mutation operation on a given chromosome.
This section presents the proposed GA for solving the multi-objective optimization problem to
maximize system reliability of a flow network, minimize the total lead-time and cost which
described in section 3, with its fuzzy linear optimization presented in section 4.the steps of this
algorithm are as follow:
6. EXPERIMENTAL RESULTS
In this section we illustrated the results of applying the proposed approach on three networks,
four nodes, six nodes and TANET (Taiwan Academic Network). The genetic parameters used in
the proposed GA are: 𝒮 = 10, ℊ = 100, 𝑔𝑐 = 0.95, 𝑔𝑚 = 0.05, 0.3 ≤ 𝛼 ≤ 0.8.
The network shown in Figure3 has four nodes and six arcs. The capacity, probability, lead-time
and cost of each component (𝑣𝑛) is shown in Table1. There are six minimal paths:
𝑚𝑝1 = {𝑎1 , 𝑎2 }, 𝑚𝑝2 = {𝑎1 , 𝑎5 , 𝑎8 }, mp3 = {a1 , a2 , a6 }, mp4 = {a1 , a2 , a7 a8 }, mp5
= {𝑎3 , 𝑎6 } 𝑎𝑛𝑑 𝑚𝑝6
= {𝑎3 , 𝑎7 , 𝑎8 }.We studied different values for T under different values of 𝛼 when d=4 as
illustrated in table 2, 3, 4, 5.where 𝑐𝑜𝑏𝑗 = 200, 𝑐0 = 250, 𝑅𝑜𝑏𝑗 = 0.99, 𝑅 = 0.9, 𝑆𝑜𝑏𝑗 , = 9, 𝑆0 =
12.
Capacity
𝒗𝒏𝒌 𝒍(𝒗𝒏𝒌 ) 𝒄(𝒗𝒏𝒌 )
0 1 2 3 4 5 6
7
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
7 0.50 0.50 0 0 0 0 0 3 20
Table 2. Optimal results founded by proposed approach to the network in fig.3, when T=6.
Table 3. Optimal results founded by proposed approach to the network in fig.3, when T=7.
Table 4.Optimal results founded by proposed approach to the network in fig.3, when T=8.
Table 5.Optimal results founded by proposed approach to the network in fig.3, when T=9.
The network has six nodes and 9 links (Fig. 4), [9].The 𝑀𝑃𝑠 are as follow:
𝑚𝑝1 = {𝑎1 , 𝑎4 , 𝑎9 }, 𝑚𝑝2 = {𝑎1 , 𝑎4 , 𝑎7 , 𝑎8 }, 𝑚𝑝3 = {𝑎1 , 𝑎5 , 𝑎8 }, 𝑚𝑝4 = {𝑎1 , 𝑎5 , 𝑎7 , 𝑎9 },
𝑚𝑝5 = {𝑎1 , 𝑎3 , 𝑎6 𝑎8 }, 𝑚𝑝6 = {𝑎1 , 𝑎3 , 𝑎6 , 𝑎7 , 𝑎9 }, 𝑚𝑝7 = {𝑎2 , 𝑎6 𝑎8 }, 𝑚𝑝8 = {𝑎2 , 𝑎6 , 𝑎7 , 𝑎9 },
𝑚𝑝9 = {𝑎2 , 𝑎3 𝑎4 , 𝑎9 }, 𝑚𝑝10 = {𝑎2 , 𝑎3 , 𝑎4 , 𝑎7 , 𝑎8 }, 𝑚𝑝11 = {𝑎2 , 𝑎3 𝑎5 , 𝑎8 }, 𝑚𝑝12 = {𝑎2 , 𝑎3 , 𝑎5 , 𝑎7 , 𝑎9 }.
8
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
We studied different values for d, T under different values for 𝛼 as illustrated in table 7, 8,9,10
.where
cobj = 450, c0 = 550, R obj = 0.99, R 0 = 0.9, Sobj = 14, S0 = 19.
a4
a1 a9
s a3 a5 a7 t
a2 a8
a6
Figure 4. The six-nodes network example
Table 6. Arc capacity, probability, lead-time, and cost for the 20 available components
Capacity
𝒗𝒏𝒌 𝒍(𝒗𝒏𝒌 ) 𝒄(𝒗𝒏𝒌 )
0 1 3 4 5 6
1 0.01 0.00 0.01 0.00 0.01 0.00 0.97 2 10
2 0.05 0.05 0.05 0.15 0.20 0.50 0 3 60
3 0.07 0.08 0.00 0.85 0 0 0 2 10
4 0.70 0.00 0.00 0.00 0.00 0.30 0 2 20
5 0.01 0.00 0.00 0.05 0.00 0.00 0.94 1 50
6 0.01 0.00 0.00 0.01 0.00 0.00 0.98 3 60
7 0.50 0.50 0 0 0 0 0 3 20
8 0.25 0.25 0.50 0 0 0 0 1 50
9 0.15 0.25 0.10 0.10 0.10 0.10 0.20 2 80
10 0.00 0.05 0.05 0.90 0 0 0 2 100
11 0.01 0.99 0 0 0 0 0 1 70
12 0.02 0.00 0.05 0.00 0.05 0.00 0.88 1 60
13 0.07 0.00 0.28 0.00 0.00 0.65 0 3 10
14 0.05 0.05 0.90 0 0 0 0 2 20
15 0.60 0.40 0 0 0 0 0 2 50
16 0.15 0.00 0.00 0.00 0.85 0 0 1 60
17 0.10 0.10 0.10 0.70 0 0 0 1 20
18 0.70 0.00 0.00 0.00 0.00 0.30 0 3 50
19 0.07 0.18 0.75 0 0 0 0 2 80
`20 0.40 0.40 0.20 0 0 0 0 3 100
Table 7.Optimal solutions founded by proposed approach to the network in fig.4, when d=6, T=7.
9
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
Table 8. Optimal solutions founded by proposed approach to the network in fig.4, when d=6, T=8.
Table 9. Optimal solutions founded by proposed approach to the network in fig.4, when d=6, T=9.
Table 10. Optimal solutions founded by proposed approach to the network in fig.4, when d=8, T=9.
In this section, we study the Taiwan Academic Network (TANET) with 30 and 33 links.The
available 80 components are shown in Table 11, [7],in addition ,we study TANET with 33 links
using different components information shown in table 20 , [38].
TANET with one source and one sink as shown in figure 5 has 6 MPs found by [7]. The 6 paths
are as follows: 𝑚𝑝1 = {𝑎1 , 𝑎2 , 𝑎3 , 𝑎4 , 𝑎5 , 𝑎6 , 𝑎7 , 𝑎8 , 𝑎9 , 𝑎10 , 𝑎11 , 𝑎12 , 𝑎13 },
𝑚𝑝2 = {𝑎1 , 𝑎2 , 𝑎21 , 𝑎15 , 𝑎16 , 𝑎17 , 𝑎19 , 𝑎20 }, 𝑚𝑝3 = {𝑎1 , 𝑎2 , 𝑎3 , 𝑎4 , 𝑎5 , 𝑎6 , 𝑎7 , 𝑎8 , 𝑎18 , 𝑎19 , 𝑎20 },
𝑚𝑝4 = {𝑎14 , 𝑎15 , 𝑎16 , 𝑎17 , 𝑎19 , 𝑎20 }, 𝑚𝑝5 = {𝑎22 , 𝑎23 , 𝑎24 , 𝑎25 , 𝑎26 , 𝑎27 , 𝑎28 },
𝑚𝑝6 = {𝑎22 , 𝑎23 , 𝑎24 , 𝑎25 , 𝑎26 , 𝑎27 , 𝑎29 , 𝑎30 }.
The corresponding lead-time for each component is randomly assigned in this article.We studied
different values for d, T under different values for 𝛼 as illustrated in table 14,15,16,17.
Where𝑐𝑜𝑏𝑗 = 1700, 𝑐0 = 2000, 𝑅𝑜𝑏𝑗 = 0.999, 𝑅0 = 0.9, , 𝑆𝑜𝑏𝑗 = 66, 𝑆0 = 80.
10
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
Capacity
𝐯𝐧𝐤 𝐥(𝐯𝐧𝐤 ) 𝐜(𝐯𝐧𝐤 )
0 1 2 3 4
11
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
12
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
Table 12. Optimal solutions founded by proposed approach to the network in fig.5, when d=4, T=16.
Table 13.Optimal solutions founded by proposed approach to the network in fig.5, when d=6, T=16.
13
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
50 28 1 32 15 54 39 2 34 72 37 14 49 57
0.5 0.996 0.998603 44 1615 5 60 67 61 23 79 33 19 30 55 27 7 3 78
38 18
79 19 1 6 42 33 38 2 72 39 54 8 49 21 10
0.6 0.987 0.997724 41 1480 31 14 80 75 3 4 56 23 60 13 55 44 15 73
34
2 28 49 31 8 54 79 39 52 18 6 50 72 33
0.7 0.999 0.998926 41 1345 14 73 38 27 61 71 3 1 44 23 13 60 15 20
19 21
9 4 1 39 54 56 31 38 80 6 72 28 19 23 34
0.8 0.999 0.998934 42 1485 57 21 44 49 27 18 30 61 73 24 79 32 71
50 14
Table 14.Optimal solutions founded by proposed approach to the network in fig.5, when d=8, T=18.
Table 15.Optimal solutions founded by proposed approach to the network in fig.5, when d=9, T=18.
TANET with two sources and two sinks shown in Figure 6, it has 14 MPs found by [38].The 14
paths are as follows:
𝑚𝑝1 = {𝑎4 , 𝑎32 , 𝑎17 , 𝑎18 , 𝑎19 , 𝑎12 , 𝑎13 , 𝑎14 , 𝑎15 , 𝑎16 }, 𝑚𝑝2 = {𝑎4 , 𝑎32 , 𝑎17 , 𝑎18 , 𝑎19 , 𝑎20 , 𝑎21 , 𝑎22 },
𝑚𝑝3 = {𝑎5 , 𝑎17 , 𝑎18 , 𝑎19 , 𝑎12 , 𝑎13 , 𝑎14 , 𝑎15 , 𝑎16 }, 𝑚𝑝4 = {𝑎5 , 𝑎17 , 𝑎18 , 𝑎19 , 𝑎20 , 𝑎21 , 𝑎22 },
14
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
𝑚𝑝5 = {𝑎6 , 𝑎7 , 𝑎8 , 𝑎9 , 𝑎10 , 𝑎11 , 𝑎12 , 𝑎13 , 𝑎14 , 𝑎15 , 𝑎16 }, 𝑚𝑝6
= {𝑎6 , 𝑎7 , 𝑎8 , 𝑎9 , 𝑎10 , 𝑎11 , 𝑎20 , 𝑎21 , 𝑎22 },
𝑚𝑝7 = {𝑎1 , 𝑎32 , 𝑎17 , 𝑎18 , 𝑎19 , 𝑎12 , 𝑎13 , 𝑎14 , 𝑎15 , 𝑎16 }, 𝑚𝑝8 = {𝑎1 , 𝑎32 , 𝑎17 , 𝑎18 , 𝑎19 , 𝑎20 , 𝑎21 , 𝑎22 },
𝑚𝑝9 = {𝑎5 , 𝑎17 , 𝑎18 , 𝑎19 , 𝑎12 , 𝑎13 , 𝑎14 , 𝑎15 , 𝑎16 }, 𝑚𝑝10 = {𝑎2 , 𝑎17 , 𝑎18 , 𝑎19 , 𝑎20 , 𝑎21 , 𝑎22 },
𝑚𝑝11 = {𝑎3 , 𝑎23 , 𝑎24 , 𝑎25 , 𝑎26 , 𝑎27 , 𝑎28 }, 𝑚𝑝12 = {𝑎3 , 𝑎23 , 𝑎24 , 𝑎25 , 𝑎26 , 𝑎27 , 𝑎29 },
𝑚𝑝13 = {𝑎3 , 𝑎23 , 𝑎24 , 𝑎25 , 𝑎26 , 𝑎27 , 𝑎33 , 𝑎30 }, 𝑚𝑝14 = {𝑎3 , 𝑎23 , 𝑎24 , 𝑎25 , 𝑎26 , 𝑎27 , 𝑎33 , 𝑎31 }.
We studied different values for d, T under different values for 𝛼as illustrated in table 19, 20, 21,
22.Where 𝑐𝑜𝑏𝑗 = 1700, 𝑐0 = 2000, 𝑅𝑜𝑏𝑗 = 0.999, 𝑅0 = 0.9, , 𝑆𝑜𝑏𝑗 = 66, 𝑆0 = 80.
Table 16. Optimal results founded by proposed approach to the network in fig.6, when d=4, T=16.
Table 17. Optimal results founded by proposed approach to the network in fig.6, when d=6, T=16.
15
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
Table 18. Optimal results founded by proposed approach to the network in in fig.6, when d=8, T=18.
Table 19: optimal results founded by proposed approach to the network in fig.6, when d=9, T=18.
As shown in section 6.3.2.TANET has 33 links and 14 MPs. We studied different values for d, T
under different values for 𝛼,when 𝛼 = 0.5,0.6,0.7 𝑎𝑛𝑑 0.8 no solutions found as illustrated in
table 21,when
𝛼 = 0.6,0.7 𝑎𝑛𝑑 0.8 no solutions found as illustrated in table 22, 23, 24.Where 𝑐𝑜𝑏𝑗 =
1700, 𝑐0 = 3000, 𝑅𝑜𝑏𝑗 = 0.999, 𝑅0 = 0.9, , 𝑆𝑜𝑏𝑗 = 66, 𝑆0 = 80.
16
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
Table 20. Component information.
Capacity
𝐯𝐧𝐤 𝐥(𝐯𝐧𝐤 𝐜(𝐯𝐧𝐤
0 1 2 3 4
1 0.0004 0.0392 0.9604 0 0 1 100
2 0.000512 0.017664 0.203136 0.778688 0 1 50
3 0.000343 0.013671 0.181629 0.804357 0 1 65
4 0.015 0.985 0 0 0 2 80
5 0.0016 0.0768 0.9216 0 0 2 70
6 0.005929 0 0.142142 0 0.851929 1 135
7 0.003 0 0.997 0 0 2 60
8 0.007225 0 0.15555 0 0.837225 1 35
9 0.005929 0 0.142142 0 0.851929 1 35
10 0.003 0.997 0 0 0 2 80
11 0.034 0.966 0 0 0 2 55
12 0.0036 0.1128 0.8836 0 0 3 40
13 0.000001 0.000297 0.029403 0.970299 0 2 110
14 0.000784 0.054432 0.944784 0 0 1 65
15 0.000225 0.02955 0.970225 0 0 1 70
16 0.095 0.905 0 0 0 3 15
17 0.005776 0.140448 0.853776 0 0 3 35
18 0.000625 0.04875 0.950625 0 0 2 75
19 0.000729 0.022113 0.223587 0.753571 0 1 40
20 0.001 0.027 0.243 0.729 0 2 35
21 0.000512 0.017664 0.203136 0.778688 0 1 45
22 0.004225 0.12155 0.874225 0 0 3 30
23 0.005929 0 0.142142 0 0.851929 1 85
24 0.003 0 0.997 0 0 2 70
25 0.000216 0.010152 0.159048 0.830584 0 3 55
26 0.034 0.966 0 0 0 2 30
27 0.000512 0.017664 0.203136 0.778688 0 2 55
28 0.000343 0.013671 0.181629 0.804357 0 1 60
29 0.001 0.027 0.243 0.729 0 3 35
30 0.0009 0.0582 0.9409 0 0 2 85
31 0.002809 0.100382 0.896809 0 0 1 60
32 0.000166375 0.008575875 0.147349125 0.843908625 0 2 70
33 0.000125 0.007125 0.135375 0.857375 0 2 80
34 0.0001 0.0198 0.9801 0 0 1 140
35 0.025 0.975 0 0 0 3 10
36 0.024 0.976 0 0 0 3 60
37 0.000125 0.007125 0.135375 0.857375 0 2 75
38 0.000110592 0.006580224 0.130507776 0.862801408 0 1 85
39 0.0001 0 0.0198 0 0.9801 1 100
40 0.001849 0 0.082302 0 0.915849 3 60
41 0.001024 0.061952 0.937024 0 0 2 60
42 0.000676 0.050648 0.948676 0 0 2 65
43 0.007921 0.162158 0.829921 0 0 4 35
44 0.000512 0.017664 0.203136 0.778688 0 2 25
45 0.001 0.027 0.243 0.729 0 5 20
46 0.097 0 0.903 0 0 4 40
47 0.000001 0.000297 0.029403 0.970299 0 3 135
48 0.022 0.978 0 0 0 2 70
49 0.000256 0 0.031488 0 0.968256 1 145
50 0.001225 0 0.06755 0 0.931225 1 70
51 0.025 0.975 0 0 0 3 70
52 0.000274625 0.011851125 0.170473875 0.817400375 0 2 65
17
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
Table 21. Optimal results founded by proposed approach to the network in fig.6, when d=4, T=16.
Table 22. Optimal results founded by proposed approach to the network in fig.6, when d=6, T=16.
Table 23. Optimal results founded by proposed approach to the network in in fig.6, when d=8, T=18.
69 8 21 33
71 59 39 53 5 11 60 78 72 8 64 44 28 14
0.4 0.438 0.991574 61 2305 30 9 26 54 31 52 56 37 20 23 2 79 1 43 74
10 33 38 13
24 12 61 53 77 40 6 21 7 52 60 48 1 3 72
0.5 0.538 0.997141 59 2465 34 8 9 38 15 23 62 59 42 54 50 29 41 32
75 55 79 20
Table 24. Optimal results founded by proposed approach to the network in fig.6, when d=9, T=18.
19
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
8. CONCLUSIONS
An approach based on GA with fuzzy optimization is presented in this paper. The presented
approach was succeeded to solve the optimal CAP problem in which each components has three
attributes; probability, cost, and lead-time. Using fuzzy membership function as fitness, the
proposed approach succeeded to find the best optimal solution with maximum system reliability,
minimum total assignment cost, and minimum total lead-time in comparison with previous
algorithms.
REFERENCES
[1] Y. K. Lin & Huang Cheng-Fu, (2013) “Stochastic Flow Network Reliability with Tolerable Error
Rate”, Quality Technology and Quantitative Management, Vol. 10, No. 1, pp. 57-73.
[2] http://etd.ohiolink.ed,Xuinying Wu,(2014),Master of Science , “Heuristic for Multi-type Component
Assignment Problems through The Birnbaum Importance”,The Faculty of The Russ College of
Engineering and Technology, Ohio University.
[3] Y. K. Lin (2010) “Reliability of k Separate Minimal Paths Under Both Time and Budget Constraints”,
IEEE Trans. Reliability, 59: 183-190.
[4] Y.K. Lin & C. T. Yeh , ( 2011) “Maximizing Network Reliability for Stochastic Transportation
Networks Under a Budget Constraint by Using a Genetic Algorithm”, International Journal of
Innovative Computing Information and Control, 7(12): 7033-50.
[5] Y.K. Lin & C.T.Yeh, ( 2011) “Multistate Components Assignment Problem with Optimal Network
Reliability Subject to Assignment Budget”, Applied Math. Comput., 217: 10074-10086.
[6] Y.K. Lin & C.T. Yeh, (2012) “Multi-objective Optimization for Stochastic Computer Networks
Using NSGA-II and TOPSIS”, European Journal of Operational Research, Vol. 218, No. 3, pp. 735-
746.
[7] Y.K. Lin & C.T.Yeh, ( 2013) “A Two-stage Approach for a Multi-objective Component Assignment
Problem for a Stochastic-flow Network”, Eng. Optimiz., 45: 265-285. DOI:
10.1080/0305215X.2012.669381.
[8] S. G. Chen, (2014) “Optimal Double-resource Assignment for The Robust Design Problem in
Multistate Computer Networks”, Applied Math. Model., 38: 263-277. DOI:
10.1016/j.apm.2013.06.020.
[9] M.R.Hassan, (2015) “Solving a Component Assignment Problem for a Stochastic Flow Network
under Lead-time Constraint”,Indian Journal of Science and Technology, Vol. 8(35), DOI:
10.17485/ijst/2015/v8i35/70455.
[10] M.R.Hassan& H.Abdou, (2018) “Multi-objective Components Assignment Problem Subject to Lead-
time Constraint”,Indian Journal of Science and Technology, Vol. 11(21), DOI:
10.17485/ijst/2018/v11i21/100080.
[11] A. Aissou, A. Daamouche & M.R.Hassan , (2019) “Optimal Components Assignment Problem for
Stochastic Flow Network “,Journal of Computer Science , DOI:10.3844/jcssp.
[12] R. E. Bellman& L. A.Zadeh, (1970) “Decision-making in a Fuzzy Environment”, Management
Science, Vol. 17, No. 4, pp. 141-164.
[13] H. J. Zimmermann, (1978) “Fuzzy Programming and Linear Programming with Several Objective
Functions, Fuzzy Sets and Systems, Vol. 1, No.1, pp. 45-56.
[14] M.K.Luhandjula , (1989)”Fuzzy Optimization: An Appraisal”, Fuzzy Sets and Systems, Vol. 30, pp.
257-282.
[15] D. Wang, (1995) ”An Inexact Approach for Linear Programming with Fuzzy Objective and
Resources”, Fuzzy Sets and Systems, Vol. 1, No. 24, pp. 261-281.
[16] J. Tang& D.wang, (1997) ”An Interactive Approach Based on a GA for a Type of Quadratic
Programming Problem with Fuzzy Objective and Resources”, Computers and Operations Research,
vol. 24, pp. 413-422.
20
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
[17] H. R. Maleki &M. Mashinchi,(2001) “A Method for Solving a Fuzzy Linear Programming”, Journal
of Applied Mathematics and Computing, DOI: 10.1007/BF02941971.
[18] P. M. Vasant, (2005) “Solving Fuzzy Linear Programming Problems with Modified S-curve
Membership Function”, International Journal of Uncertainty, Fuzziness and knowledge-based
Systems, Vol. 13, No. 01, pp. 97-109.
[19] H. R. Maleki & Mashaallah Mashinchi, (2008) “A New Method for Solving Fuzzy Linear
Programming by Solving Linear Programming”, Applied Mathematical Sciences, Vol. 2,No. 50,pp.
2473 – 2480.
[20] C.Veeramani, C.Duraisamy & A.Nagoorgani, (2011) “Solving Fuzzy Multi-objective Linear
Programming Problems with Linear Membership Functions”, Australian Journal of Basic and Applied
Sciences, 5(8): 1163-1171.
[21] J. L. Verdegay & B. Kheirfam, (2013) “Optimization and Reoptimization in Fuzzy Linear
Programming Problems”, The 8th Conference of The European Society for Fuzzy Logic and
Technology (EUSFLAT), DOI: 10.2991/eusflat.2013.80.
[22] M.Kiruthiga & C.Loganathan, (2015)“Fuzzy Multi-objective Linear Programming Problem Using
Fuzzy Programming Model”, International Journal of Science, Engineering and Technology Research,
Vol. 4, Issue 7.
[23] S. K. Das, (2017)” Modified Method for Solving Fully Fuzzy Linear Programming Problem with
Triangular Fuzzy Numbers” ,International Journal of Research in Industrial Engineering,Vol. 6, No. 4
(2017) 293–311,DOI: 10.22105/riej.2017.101594.1024.
[24] S. Kumar & T. Mandal, (2017) ”A new model for solving fuzzy linear fractional programming
problem with ranking function”, Journal of Applied Research on Industrial Engineering,Vol. 4, No. 2
(2017) 89–96.
[25] S. H. Nasseri & H. Zavieh, (2018) “A Multi-objective Method for Solving Fuzzy LinearProgramming
Based on Semi-infinite Model”,Fuzzy Information and Engineering,VOL. 10, NO. 1, 91–98.
[26] C.Malathi1 & P.Umadevi,(2018) “A New Procedure for Solving Linear Programming Problems in An
Intuitionistic Fuzzy Environment”, International Conference on Applied and Computational
Mathematics, IOP Conf. Series: Journal of Physics: Conf. Series 1139 (2018) 012079,
doi:10.1088/1742-6596/1139/1/012079.
[27] D. S. Dinagar & M. M. Jeyavuthin,(2018) “Fully Fuzzy Integer Linear Programming Problems
UnderRobust Ranking Techniques”,International Journal of Mathematics And its Applications,
6(3)(2018), 19-25.
[28] Z. Gong,W. Zhao&K. Liu,(2018) “A Straightforward Approach for Solving Fully Fuzzy Linear
Programming Problem with LR-Type Fuzzy Numbers”,Journal of the Operations Research Society of
Japan,Vol. 61, No. 2, pp. 172-185.
[29] S. M. Ingle,(2019) “Solving FFLPP Problem with Hexagonal Fuzzy Numbersby New Ranking
Method”,International Journal of Applied Engineering Research,Vol. 14, No. 1,pp.97-101.
[30] M. Ranjbar &S. Effati,(2019) “Symmetric and right-hand-side hesitant fuzzy linear programming”,
IEEE Transactions on Fuzzy Systems,DOI: 10.1109/TFUZZ.2019.2902109.
[31] A. Kabiraj, P. K. Nayak & S. Raha, (2019) “Solving Intuitionistic Fuzzy Linear Pogramming
Problem”, International Journal of Intelligence Science, DOI: 10.4236/ijis.2019.91003.
[32] D. Wang, (1996) ”Modeling and Optimization for a Type of Fuzzy Nonlinear Programming Problems
in Manufacture Systems”, Proceeding of IEEE Conference on Decision and Control, vol. 4, pp. 4401-
4405.
[33] M. Mutingi, (2014),” System Reliability Optimization: A Fuzzy Multi-objective Genetic Algorithm
Approach”, Maintenance and Reliability, 16 (3): 400–406.
[34] H. Y. Chang, Y. J. Tzang, C. H. Tzang & C. Y. Huang, (2015) “An Application of Fuzzy Multi-
objective Linear Programming for Components Design of Games or Animated Characters”, First
International Conference on Computational Intelligence Theory, Systems and Applications ,DOI
10.1109/CCITSA.2015.39.
[35] D. Stephen Dinagar & S. Kamalanathan, (2017) ”International Journal of Applications of Fuzzy Sets
and Artificial Intelligence”,Vol. 7 (2017), 281-292.
21
International Journal of Mobile Network Communications & Telematics ( IJMNCT), Vol.9, No.3, June 2019
AUTHORS
Heba Hamdy Ahmed is a Demonstrator in Computer Science Branch, Department of Mathematics,
Faculty of Science, Aswan University, Aswan, Egypt.
Mohamed Eid Mohamedis a lecture in Computer Science Branch, Department of Mathematics, Faculty of
Science, Aswan University, Aswan, Egypt.
Mosa khalifa Ahmed is an Assistant Professor in Department of Mathematics, Faculty of Science, South
Valley University, Qena, Egypt.
22